Role of diet

Surprisingly, the relationship between diet and obesity in preschool children is poorly researched but is likely to involve interaction between appetite, dietary composition, feeding or meal patterns and parental feeding influences⁽Reference Newby¹⁴⁾. Research into the dietary determinants of preschool obesity has focused either on the role of total energy intake or the contribution of individual macronutrients⁽Reference Gunther, Buyken and Kroke^15–Reference Welsh, Cogswell and Rogers¹⁸⁾.

Reports from studies investigating associations between energy intake and obesity in preschool children are conflicting. Two studies have found that high energy intake is associated with increased obesity risk⁽Reference Stunkard, Berkowitz and Schoeller^19, Reference Cowin and Emmett²⁰⁾, whilst surprisingly a randomised controlled trial has found that a lower energy intake during the preschool years is associated with increased risk of later obesity⁽Reference Niinikoski, Viikari and Ronnemaa^21, Reference Hakanen, Lagstrom and Kaitosaari²²⁾. This discrepancy may be a consequence of difficulties in assessing energy intake, which can be confounded by factors such as body weight, metabolism and physical activity level. For example, when assessing energy intake studies do not always adjust for body size and physical activity level and few measure metabolisable energy intake using the ‘gold standard’ doubly-labelled-water method. Evidence from the USA suggests that increased energy intake rather than reduced energy expenditure is the major driver of obesity⁽Reference Swinburn, Sacks and Ravussin²³⁾. A meta-analysis of data from seven studies including 963 children aged 4–18 years has found that total energy intake is the greatest contributor to increases in body weight⁽Reference Swinburn, Jolley and Kremer²⁴⁾. More data from preschool children using the doubly-labelled-water technique to assess energy expenditure are needed before firm conclusions can be made.

Several studies have investigated the influence of macronutrient intake and the development of obesity in preschool children. The evidence is most consistent for an association between higher protein intake and increased obesity risk⁽Reference Gunther, Buyken and Kroke^15–Reference Rolland-Cachera, Deheeger and Akrout^17, Reference Stunkard, Berkowitz and Schoeller^19, Reference Verduci, Radaelli and Stival^25, Reference Scaglioni, Agostoni and Notaris²⁶⁾. For instance, one study has shown that higher protein intake in the preschool years (14–15% total energy intake v. 11–12% total energy intake) more than doubles the risk of being overweight (OR 2·34 (95% CI 1·1, 5·0))⁽Reference Gunther, Buyken and Kroke¹⁶⁾. Higher protein intake (16·6% energy intake v. 14·9% energy intake) is also associated with earlier age at adiposity rebound (<4 years of age, P<0·01), the point at which BMI begins to increase after reaching its childhood nadir (usually between 4 and 8 years of age)⁽Reference Rolland-Cachera, Deheeger and Akrout¹⁷⁾, which itself is associated with an increased risk of later obesity⁽Reference Cole^27, Reference Rolland-Cachera²⁸⁾.

Evidence for an influence of fat and carbohydrate intake on obesity risk is less convincing. Two studies suggest that higher fat intake increases the risk of obesity⁽Reference Hakanen, Lagstrom and Kaitosaari^22, Reference Magarey, Daniels and Boulton²⁹⁾ whilst another study has found no such association⁽Reference Stunkard, Berkowitz and Schoeller¹⁹⁾. Most studies investigating carbohydrate intake have focused on sweetened beverages as the source of carbohydrate and have reported conflicting results. For instance, in a recent study of >10 000 US preschool children consumption of more than one sweetened drink daily has been shown to double the risk of becoming overweight⁽Reference Welsh, Cogswell and Rogers¹⁸⁾. In contrast, other studies have not found an association between sweetened beverage consumption and risk of overweight⁽Reference Scaglioni, Agostoni and Notaris^26, Reference Magarey, Daniels and Boulton^29–Reference Skinner, Bounds and Carruth³¹⁾. Thus, although approximately half 18-month-old children consume sweetened drinks daily in the UK, it is not known whether this factor increases the risk of obesity⁽Reference Northstone, Rogers and Emmett³²⁾.

Overall, there is limited evidence to support the role of diet alone in the development of preschool obesity, although there are data to support the hypothesis that a high protein intake during infancy and early childhood increases the tendency to obesity.

Role of physical activity

Evidence suggests that increased physical activity reduces the risk of obesity in preschool children⁽Reference Mo-suwan, Pongprapai and Junjana^33–Reference Berkowitz, Agras and Korner³⁷⁾. One cross-sectional study has shown that children with lower physical activity are approximately four times (3·8 (95% CI 1·4, 10·6) times) more likely to have increased body fat⁽Reference Moore, Nguyen and Rothman³⁸⁾. The effect size of physical activity on obesity may be considerable. For example, in Thailand preschool children reported to have low exercise levels have been shown to have >2-fold risk of obesity compared with children with a normal exercise level⁽Reference Mo-suwan and Geater³⁹⁾.

The intensity of physical activity may be particularly important in influencing obesity risk. For example, one study has shown that overweight children spend approximately 9 min less time (22·9 min v. 32·1 min, P<0·05) in vigorous physical activity daily compared with normal-weight children⁽Reference Metallinos-Katsaras, Freedson and Fulton³⁴⁾. Another study has found that children in the highest quartile for vigorous activity at 4–6 years of age are in the lowest quartile for body fat percentage at 3 year follow-up. No differences in body fat percentage were seen in association with low or moderate physical activity⁽Reference Janz, Burns and Levy⁴⁰⁾. Thus, even small differences in vigorous physical activity could contribute to the development of obesity.

There are only two studies of physical activity interventions in preschool children, neither of which report a positive effect. An intervention to increase physical activity and reduce BMI in Scottish nursery school children was unsuccessful⁽Reference Reilly, Kelly and Montgomery¹²⁾, whilst an intervention in Thailand has shown a trend (P=0·06) towards reduced obesity prevalence in the intervention group⁽Reference Mo-suwan, Pongprapai and Junjana³³⁾.

Although the extent of the contribution of physical activity to preschool obesity is unclear, adopting a physically-active lifestyle is crucial for long-term health promotion. However, despite increasing awareness, most preschool children do not achieve the recommended 60 min moderate–vigorous physical activity daily⁽Reference Finn, Johannsen and Specker^41–Reference Cardon and De Bourdeaudhuij⁴³⁾.

Developmental factors

Recently, developmental factors in early life such as nutrition and the pattern of early growth have been shown to influence the later risk of obesity. For instance, rapid growth during infancy is strongly related to obesity risk⁽Reference Baird, Fisher and Lucas^44–Reference Ong and Loos⁴⁶⁾, an effect largely attributed to nutrition. Indeed, four systematic reviews now provide strong evidence for a protective effect of breast-feeding on risk of obesity⁽Reference Arenz, Ruckerl and Koletzko^47–Reference Harder, Bergmann and Kallischnigg⁵⁰⁾, possibly as a result of differences in nutritional composition between human and formula milk. For example, early protein intake has been reported to be 70% greater in formula-fed infants than in breast-fed infants, which could increase the rate of infant growth and hence the risk of later obesity. Evidence to support this notion has emerged from a large trial of >1000 infants in which infants randomised to a higher protein diet during the first year of life were found to have greater weight gain in infancy and greater BMI at 2 years of age⁽Reference Koletzko, von Kries and Monasterolo⁵¹⁾. Overfeeding in early postnatal life may therefore programme obesity development later in life.

The mechanisms for the effect of nutrition and growth in infancy on the development of obesity are poorly understood. Evidence suggests that programming of energy balance is central to energy regulation and hence the development of obesity, although there are few data in human subjects to support this hypothesis. However, animal studies suggest that neuro-endocrine systems controlling energy balance develop during the perinatal period, probably controlled by the action of hormones such as insulin and leptin on the central nervous system. Environmental factors in perinatal life, including nutrition, could therefore affect the development of appetite and hence obesity risk. For example, in mice growth restriction as a result of undernutrition in utero followed by postnatal overfeeding leads to increased body weight and adiposity in adult life⁽Reference Vickers, Krechowec and Breier^52, Reference Yura, Itoh and Sagawa⁵³⁾.

The role of developmental nutritional factors beyond infancy, when diet is no longer predominantly milk based, is largely unexplored⁽Reference Parsons, Power and Logan¹¹⁾. However, the preschool period is a pivotal time during which long-term dietary habits are established, with potential lifelong effects on obesity⁽Reference Nicklas, Farris and Smoak⁵⁴⁾. For instance, appetite in preschool children strongly influences later adiposity⁽Reference Wardle, Carnell and Haworth⁵⁵⁾. Development of particular food preferences and consumption of a greater diversity of energy-dense foods could therefore influence the development of obesity⁽Reference Nicklaus⁵⁶⁾.

Food preference is a strong determinant of dietary diversity. Both food preference and diversity track from the first year of life⁽Reference Skinner, Carruth and Houck⁵⁷⁾ and therefore dietary behaviours adopted in early childhood are likely to persist throughout life⁽Reference Nicklaus⁵⁶⁾. A greater variety of food choice is associated with increased energy intake and a greater risk of obesity in several⁽Reference Hetherington, Foster and Newman^58, Reference Rolls, Rowe and Rolls⁵⁹⁾ but not all studies⁽Reference Nicklaus, Boggio and Chabanet⁶⁰⁾, possibly as a consequence of more energy-dense foods being consumed in individuals who have a diverse diet.

Societal changes

Whilst individual behaviours are influential in the development of obesity, these behaviours are strongly influenced by societal factors. The recent dramatic increase in obesity is partly attributable to societal changes that encompass educational, financial and recreational aspects⁽Reference Law, Power and Graham⁹⁾. For example, improvements in food manufacturing and marketing have resulted in increased availability of energy-dense foods⁽Reference Hawkins and Law⁶¹⁾. Eating habits have also changed markedly. More food is now eaten outside the home, with energy-dense ‘convenience foods’ more commonly consumed⁽Reference Nielsen and Popkin⁶²⁾. At the same time, increased use of technology such as lifts, moving walkways and cars has reduced the time spent in physical activity. Schoolchildren are less likely to walk to school⁽Reference Sturm⁶³⁾, while televisions and personal computers may encourage sedentary behaviours⁽Reference Sturm⁶⁴⁾. However, the role of these behaviours is controversial and despite the obesity epidemic it has been suggested that there has been no increase in television viewing in the USA⁽Reference Sturm⁶⁴⁾.

Changes in diet

Dietary patterns among children have changed considerably over the past 30 years⁽Reference Hawkins and Law⁶¹⁾. National surveys from the USA and UK suggest that preschool children's diets do not comply with recommendations for fruit and vegetable intake but exceed the recommended intake for fat and refined carbohydrate⁽Reference Cowin and Emmett^65–Reference Gregory, Collins and Davies⁶⁷⁾. Reports on energy intake of preschool children from national surveys are conflicting. In the UK energy intake has remained static or declined during the course of the obesity epidemic⁽Reference Prynne, Paul and Price⁶⁸⁾. For instance, the energy intake of 4-year-old children in the UK has been reported to have declined from 6046 kJ (1445 kcal)/d in 1950 to 5138 kJ (1228 kcal)/d in 1992⁽Reference Prynne, Paul and Price⁶⁸⁾. However, in the USA energy intakes of preschool children appear to have significantly increased from 5812 kJ (1389 kcal)/d in 1977 to 6519 kJ (1558 kcal)/d in 1994, possibly because of increases in portion size and increased consumption of food outside the home⁽Reference Kranz, Siega-Riz and Herring⁶⁹⁾. This discrepancy may be a result of difficulties in assessing dietary intake in this age-group, which is particularly problematic for foods consumed outside the home. In the USA, for instance, a decrease has been reported in food consumed at home, from 76% total energy during the 1970s to 69% total energy in the 1990s, together with large increases in the consumption of energy-dense foods including salty snacks, pizza and soft drinks⁽Reference Nielsen, Siega-Riz and Popkin⁷⁰⁾.

Changes in physical activity

Societal changes mean that many children spend more time in childcare facilities, where structured activities have displaced time spent in free play. For example, American preschool children spend less time playing, approximately 8·5 h per week less in 1997 compared with 1981⁽Reference Sturm⁶⁴⁾. American preschool children also spend insufficient time in recreational activity⁽Reference Datar and Sturm⁷¹⁾. One survey (in 1998) has reported that only 16% of children take part in daily physical education sessions, with 13% participating less than once per week, often for <30 min each day⁽Reference Sturm⁶³⁾.

Increased time spent in sedentary activities may contribute to positive energy balance and the development of obesity. In the USA many children exceed recommendations for a maximum of 2 h television viewing per d⁽Reference Taveras, Rifas-Shiman and Belfort^72, Reference Vandewater, Rideout and Wartella⁷³⁾. Recent increases in the use of media sources such as videos and computer games are likely to have increased the time spent in sedentary activity and ⩽30–40% of American children aged <6 years watch videos and play computer games daily⁽Reference Vandewater, Rideout and Wartella⁷³⁾. Time spent reading or being read to has also doubled in preschool children between 1981 and 1997, probably as a result of increased pressure to excel academically⁽Reference Sturm⁶⁴⁾.

In summary, secular trends at the societal and individual level have led to changes in dietary and physical activity behaviours in preschool children that have increased the risk of obesity. Interventions to redress these changes are therefore urgently sought.

Physical activity interventions

A study from Thailand in 292 preschool children randomised to either an exercise or control group has found a trend towards lower obesity prevalence in the intervention group (8·8% v. 9·4% in controls, P=0·06)⁽Reference Mo-suwan, Pongprapai and Junjana³³⁾. The second study of 545 Scottish nursery-school children has reported no significant effect on BMI at either 6 or 12 months after the intervention⁽Reference Reilly, Kelly and Montgomery¹²⁾.

Combined dietary and physical activity interventions

Three controlled trials of interventions in preschool children have combined dietary advice and physical activity⁽Reference Harvey-Berino and Rourke^76–Reference Fitzgibbon, Stolley and Schiffer⁷⁸⁾. Only one of these trials has reported success, with lower BMI z-score 2 years after intervention (difference −0·54 (95% CI –0·98, −0·10) kg/m², P=0·02)⁽Reference Fitzgibbon, Stolley and Schiffer⁷⁸⁾.

Behavioural therapy

Previous interventions have all highlighted the importance of behavioural therapies that include theoretical frameworks to address eating behaviours and the factors that influence them⁽Reference Glanz, Basil and Maibach^79–Reference Baranowski⁸¹⁾. Assessment of carers' and children's motivation to change adverse behaviours in general⁽Reference Prochaska and Velicer⁸²⁾ (and eating behaviours in particular⁽Reference Van Strien, Frijters and Bergers⁸³⁾) is important. Although family involvement is critical to the success and sustainability of obesity prevention programmes in children, only 3% of such programmes have included a family or home-based component⁽Reference Flynn, McNeil and Maloff⁷⁾.

Overall, only one intervention to prevent obesity in early childhood (age 1–5 years) has shown long-term benefits⁽Reference Fitzgibbon, Stolley and Schiffer⁷⁸⁾. This US-based programme combined dietary and physical activity intervention and included a behavioural theoretical framework. However, the validity and generalisability of this study to a UK population is unknown, and there appear to be no interventions in the UK designed to prevent obesity in preschool children that meet all the National Institute for Health and Clinical Excellence recommendations⁽⁷⁴⁾.

Recently, a new method has been described for building and evaluating internet-based health interventions⁽Reference Collins, Murphy and Strecher⁸⁴⁾ that could equally be applied to the development of obesity interventions. This strategy employs a multiphase optimisation design that requires: a screening phase in which candidate components for inclusion in an intervention are identified; a refining phase in which selected components are developed and their optimal proportion within the intervention is assessed; a confirming phase in which the optimised intervention is evaluated in a randomised controlled trial. This strategy may help in the design of successful interventions for childhood obesity and has been used in the Trim Tots obesity intervention.